AI

GeoCR: Learning a Generalist Cloud Removal Prior from Heterogeneous Observations

Read full article at source ↗

Based on a feed excerpt; full article content has not been acquired. Source evidence retrieved 06/10/2026, 14:18:24 UTC.

Briefings

GeoCR is a generalist cloud removal model that unifies RGB-only and multispectral removal from heterogeneous observations using a single network. It achieves the best FID and DISTS on full-band SEN12MS-CR and RGB-only CUHK-CR2, outperforming existing methods. The model uses a shared latent interface with frozen encoder-decoder trunks and supports LoRA adaptation for efficiency.

Key points

• GeoCR unifies RGB-only and multispectral cloud removal from single or multi-temporal observations using a single network architecture.
• The model achieves the best FID and DISTS scores on full-band SEN12MS-CR and RGB-only CUHK-CR2, outperforming existing methods.

Context

Cloud removal methods are typically specialized to individual datasets and input configurations, limiting reuse across sensors, spectral bands, and observation settings. GeoCR addresses this by learning a shared cloud removal prior across diverse settings.

Limitations

Evaluation is limited to test splits of contributing datasets under full-band and RGB-only settings. The study does not report results on SAR guidance or multi-temporal settings, and the model's performance on other heterogeneous configurations remains unverified.

Generated by Bittrees-hosted qwen35-2b sha256:aaf42c8b7c3c on 21/10/2026, 05:18:55 UTC. Based on publisher-provided feed evidence; may not cover the full original article. Check the source for details.

View archived briefing on IPFS
bafybeihgypmmzbcn5rnzp6y2moklpqd7gbpocgozsg2gjazfpdb4iyge2u